You evaluate this by asking new patients where they heard about your practice, testing the exact questions a newly diagnosed patient or caregiver would type into ChatGPT, Gemini, Perplexity, and Google AI Overviews, and watching whether your practice's name, location, and specialty show up in those answers. If your name appears consistently and new patients mention seeing it, AI search is contributing to your pipeline. If it never appears, patients relying on those tools are likely being routed to a competing practice instead.
Asking the engines the questions your patients would ask
The most direct way to check your visibility is to type patient-style questions into each AI search tool yourself and read the answers as if you were a caregiver who just got a diagnosis. Questions like "oncologist near me who treats multiple myeloma" or "hematology oncology practice in your city that takes new patients" reveal exactly what these tools surface. Run the same question across ChatGPT, Gemini, Perplexity, and Google AI Overviews (the summarized answer box that appears above traditional search results), since each pulls from different sources and may produce a different answer.
Write down the actual phrasing a patient would use rather than clinical terminology. Test variations: general questions ("best oncologist in your city"), condition-specific questions ("hematologist for anemia treatment near me"), and logistics questions ("oncology practice accepting new Medicare patients"). Each variation tests a different part of how AI tools decide which practice to mention.
What to note when your practice does or does not appear
When your practice appears in an AI-generated answer, record which tool produced it, the exact question asked, and what information the tool included about you, such as accepted insurance, subspecialties, or location. When your practice does not appear, note which competing practices did and what those practices' online presence includes that yours might not, such as more detailed service pages, patient reviews, or clearer descriptions of conditions treated.
This comparison matters because AI tools generate answers by pulling from what is publicly available and clearly structured about a practice, including your website content, review profiles, and directory listings. A practice that never appears usually has thin or inconsistent information across those sources. A practice that appears sometimes but not for every relevant question may have strong visibility for general oncology terms but nothing that signals expertise in specific conditions like lymphoma, sickle cell disease, or a particular treatment approach. Tracking both outcomes over time shows whether gaps are closing or staying put.
Signs an AI answer influenced a booking
A new patient mentioning that they "found you through ChatGPT" or "asked Google's AI about oncologists near me" during their intake call is the most direct signal, so front-desk staff should be trained to ask and log the source using the same categories as other referral tracking. Beyond direct mentions, watch for patients who arrive already familiar with specific details about your practice, such as a subspecialty focus or a particular physician's background, that they could only have gotten from a detailed AI-generated summary rather than a basic listing.
Another signal is a shift in how new patients describe their search process. If patients increasingly describe asking a question and getting a direct answer rather than browsing a list of links, that describes how AI search tools work and suggests those tools played a role even if the patient does not name the tool by brand. Comparing the phrasing patients use during intake against the phrasing you tested in the engines can confirm whether the questions you checked match the questions patients are actually asking.
Reviewing results on a regular schedule
Checking AI search visibility once and moving on is not useful, because these tools update their answers as the underlying information they pull from changes, and a practice's own website, reviews, and directory listings also change over time. Set a fixed recurring schedule, such as monthly, to re-run the same set of patient questions across each AI tool and compare results against your previous check. Keep a simple running log rather than relying on memory.
Track three things at each review: whether your practice appears for the same questions it appeared for last time, whether any new competing practices have started appearing, and whether the information included about your practice in the answer has changed or gone stale, such as an outdated address or a physician who has since left the practice. A consistent review schedule turns this from a one-time curiosity into an ongoing measure of whether your online presence is keeping pace with how patients now search.
The next step that matters more than the rest
Start the intake-source question this month: have every front-desk and scheduling staff member ask each new patient, in plain language, where they heard about the practice, and specifically whether they used ChatGPT, Gemini, Perplexity, or a Google AI summary during their search. This single habit outranks every other evaluation step because it turns AI search from something you guess about into something you measure with real patient behavior. Testing questions yourself tells you what the engines say; asking patients directly tells you whether that answer actually brought someone through your door, and that is the only number that determines whether the effort to improve your visibility is worth continuing.